It is still controversial, whether moderately high haematocrit (Ht) and haemoglobin (Hb) values are risk factors for coronary heart disease. Using the computerized data-system of the Periodical Medical Examination (PME) of Phillips' International Electrical Company, a case-control study was carried out. Cases were male workers (n = 104, from 50 to 60 years of age) who had suffered a first, non-fatal myocardial infarction, and who had had a PME prior (on the average 16 months) to the occurrence of infarction. For each case two age-matched healthy controls were selected from the PME-attendancy list (= 208). For each subject information was abstracted from the PME-records about haematologic parameters and covariates (smoking, cholesterol, blood pressure, pulse-rate, weight, height, FEV5, consumption of antihypertensive agents). After dichotomizing the haematocrit and haemoglobin values at their whole sample means (0.46 l/l and 9.7 mmol/l respectively) in "low" (lower than or equal to the mean) and "high" (greater than the mean), crude odds-ratio's of 2.7 (95% CI: 1.6-4.6) and 2.1 (95% CI: 1.2-3.6) were found for Ht and Hb respectively, when comparing "high" with "low" levels. The associations between Ht and Hb, and the occurrence of myocardial infarction were still present after controlling for covariates using multiple logistic regression models, entering the continuous variables with their exact values. After adjustment, mean corpuscular volume (MCV) also appeared to be correlated with infarction. Our results confirm the hypothesis that moderately high haematocrit and--to a lesser extent--haemoglobin and MCV-values are risk factors for the occurrence of myocardial infarction.